diff --git a/crates/admin/src/cluster_controller/cluster_state_refresher.rs b/crates/admin/src/cluster_controller/cluster_state_refresher.rs
index 792cf602b..05b8f0557 100644
--- a/crates/admin/src/cluster_controller/cluster_state_refresher.rs
+++ b/crates/admin/src/cluster_controller/cluster_state_refresher.rs
@@ -201,7 +201,7 @@ impl<T: TransportConnect> ClusterStateRefresher<T> {
                         // todo: implement a more robust failure detector
                         // This is a naive mechanism for failure detection and is just a stop-gap measure.
                         // A single connection error or timeout will cause a node to be marked as dead.
-                        trace!("Node {node_id} is marked dead {node_id}: {err}");
+                        trace!("Node {node_id} is marked dead: {err}");
                         let last_seen_alive = last_state.nodes.get(&node_id.as_plain()).and_then(
                             |state| match state {
                                 NodeState::Alive(AliveNode {
diff --git a/crates/admin/src/cluster_controller/observed_cluster_state.rs b/crates/admin/src/cluster_controller/observed_cluster_state.rs
index 260b9f569..05104b1b3 100644
--- a/crates/admin/src/cluster_controller/observed_cluster_state.rs
+++ b/crates/admin/src/cluster_controller/observed_cluster_state.rs
@@ -44,6 +44,10 @@ impl ObservedClusterState {
         self.update_partitions(cluster_state);
     }
 
+    /// Update observed cluster state with given [`ClusterState`]
+    /// Nodes in [`NodeState::Suspect`] state are treated as [`NodeState::Alive`].
+    /// This means that their `last know` state will not be cleared up yet
+    /// until they are marked as dead by the [`ClusterState`].
     fn update_nodes(&mut self, cluster_state: &ClusterState) {
         for (node_id, node_state) in &cluster_state.nodes {
             match node_state {